8/2/2019 2003 62 Gaia Conjectures
1/25
THE GAIA HYPOTHESIS: CONJECTURES AND REFUTATIONS
JAMES W. KIRCHNER
Department of Earth and Planetary Science, University of California, Berkeley,
CA 94720-4767, U.S.A.
E-mail: [email protected]
Abstract. The uncertainties surrounding global climate change provide ample evidence, if any
were necessary, of the need for a whole-system view of the Earth. Arguably the most visible
and controversial attempt to understand Earth as a system has been Lovelocks Gaia theory. Gaia
has been a fruitful hypothesis generator, and has prompted many intriguing conjectures about how
biological processes might contribute to planetary-scale regulation of atmospheric chemistry and
climate. In many important cases, however, these conjectures are refuted by the available data. For
example, Gaia theory predicts that the composition of the atmosphere should be tightly regulated bybiological processes, but rates of carbon uptake into the biosphere have accelerated by only about 2%
in response to the 35% rise in atmospheric CO2 since pre-industrial times. Gaia theory would predict
that atmospheric CO2 should be more sensitively regulated by terrestrial ecosystem uptake (which
is biologically mediated) than by ocean uptake (which is primarily abiotic), but both processes are
about equally insensitive to atmospheric CO2 levels. Gaia theory predicts that biological feedbacks
should make the Earth system less sensitive to perturbation, but the best available data suggest that
the net effect of biologically mediated feedbacks will be to amplify, not reduce, the Earth systems
sensitivity to anthropogenic climate change. Gaia theory predicts that biological by-products in the
atmosphere should act as planetary climate regulators, but the Vostok ice core indicates that CO2,
CH4, and dimethyl sulfide all biological by-products function to make the Earth warmer when it
is warm, and colder when it is cold. Gaia theory predicts that biological feedbacks should regulate
Earths climate over the long term, but peaks in paleotemperature correspond to peaks in paleo-CO2in records stretching back to the Permian; thus if CO2 is biologically regulated as part of a global
thermostat, that thermostat has been hooked up backwards for at least the past 300 million years.
Gaia theory predicts that organisms alter their environment to their own benefit, but throughout most
of the surface ocean (comprising more than half of the globe), nutrient depletion by plankton has
almost created a biological desert, and is kept in check only by the nutrient starvation of the plankton
themselves. Lastly, where organisms enhance their environment for themselves, they create positive
feedback; thus Gaia theorys two central principles first, that organisms stabilize their environment,
and second, that organisms alter their environment in ways that benefit them are mutually incon-
sistent with one another. These examples suggest that the further development of Gaia theory will
require more deliberate comparison of theory and data.
1. Introduction
My title is a whimsical tip of the hat to Poppers (1963) classic collection of essays,
in which he puts forth the view that science progresses by an iterative process of
conjecture and refutation, as theories are speculatively proposed, tested against
observations, revised, and tested again. With the benefit of hindsight one can see a
Climatic Change 58: 2145, 2003.
2003 Kluwer Academic Publishers. Printed in the Netherlands.
8/2/2019 2003 62 Gaia Conjectures
2/25
22 JAMES W. KIRCHNER
similar process underway with Gaia, although in my view there has been rather a
lot of conjecture, and rather too little willingness to recognize that these conjectures
are often refuted by the available data. I emphasize that in spite of this, I believe
that Gaia has been fruitful as a metaphor and a hypothesis generator, and I have
consistently said so in print (Kirchner, 1989, 1991, 2002).
In this contribution I have three objectives. The first is to address specific points
raised by Lenton and Wilkinson (2003) in their response to my recent commentary
(Kirchner, 2002) and the commentary by Volk (2002). The second is to use Lenton
and Wilkinsons comments as a springboard to expand and clarify my views on
Gaia and the Earth system. The third is to outline new directions for Gaia theory, in
the hope that it can contribute to the further development of Earth system science.
Particularly because I have been a vocal critic of the Gaia hypothesis, it is impor-
tant to make clear that I substantially agree with many of its central themes. I agree
with Gaia, in the sense that I think that no view of the Earth system can be complete
or even approximately correct without accounting for the pervasive influence of
biological processes on Earths surface chemistry and climate. I agree that couplingbetween the atmosphere and biosphere should naturally give rise to feedback. I also
agree that coupled feedback systems can exhibit emergent behaviors, including
self-regulation. And most importantly, I agree that it is essential to understand the
Earth system as a system, rather than as a set of disconnected components.
At a mechanistic level, I think that Gaias supporters and I largely agree about
how the Earth system works. I think we disagree, however, about how to character-
ize the behavior of that system. Specifically, I do not think that the data warrant the
generalizations that Gaias supporters have put forward, to the effect that biolog-
ical feedbacks generally enhance the self-regulation of the atmosphere-biosphere
system, or that organisms generally alter their environment to their own benefit.
I think that given the currently available data, we need to seriously consider thepossibility that in many cases, Earths climate has remained relatively stable in
spite of biologically mediated feedbacks rather than because of them.
2. Should One Generalize about Gaia?
One of my tasks in this contribution is to respond to points raised by Lenton
and Wilkinson (2003) hereafter denoted as LW where I believe a response
is required. LW caricature Volks recent commentary on Gaia (Volk, 2002), saying
Volk criticizes Kleidon and Lenton for giving only a finite number of examples
to support their arguments. In fact, Volk criticized Kleidon for giving only two
examples to support the sweeping generalization that life has a strong tendency to
affect its environment in a way that enhances the overall benefit; Volk then gave
two counter-examples to show how shaky this generalization is. And Volk did not
criticize Lenton for giving only a finite number of examples at all; instead he
questioned whether Lentons treatment was even-handed, asking, how can we be
8/2/2019 2003 62 Gaia Conjectures
3/25
GAIA: CONJECTURES AND REFUTATIONS 23
sure that Lenton is not giving more weight to evidence that fits his pre-selected con-
clusions? This is a fair question in the context of the past 30 years, during which
Gaias proponents have vigorously advanced the view that life stabilizes Earths
climate, while largely ignoring the substantial body of evidence that biological
feedbacks also can (and do) destabilize Earths climate. It is also a fair question
in the context of Lenton (2002), despite LWs assertion that an effort was made
to objectively balance available evidence. While Lenton clearly acknowledges the
theoretical possibility of biologically mediated positive feedbacks, in 14 pages he
makes only a few passing references to the biological feedbacks that actually do
destabilize Earths climate.
Later in the same paragraph, LW say, Volk does not level the same criticisms at
his own attempts to generalise or those of Kirchner. To make progress, criteria for
what constitutes a reasonable argument must be applied in a consistent manner.
This kind of indirect accusation is a debaters trick: with no supporting evidence of
any kind, it simultaneously insinuates that Volk and I have indulged in generaliza-
tions similar to Kleidons, and accuses Volk of applying a double standard. Whereis the unwarranted generalization or the generalization of any sort in Volks
paper or my own?
Indeed, I have been arguing for more than a decade that the quest for grand
generalizations is part of the problem, because it can blind us to the diversity of
feedback processes in the Earth system (Kirchner, 1989). I certainly agree with
LW that, We should not expect there to be universal truths about the behavior
of such a complex system . . . Life is not always going to enhance gross primary
productivity . . . or any other metric of the system. But I disagree with their view
that What matters is the balance of examples: do they suggest a tendency in one
direction? In my view, what matters is not the balance of examples, what matters
is how the Earth system works, in all its intriguing diversity and complexity. To beuseful, generalizations need to be almost universally true, and this is unlikely given
the great diversity of life (and environments) on Earth. Even if there is a tendency
one way or another, the exceptions will make any generalization misleading as a
guide to how the system works.
This is not to say that generalizations are not useful as hypotheses, and therefore
as spurs to investigation. It makes perfect sense to put forward the hypothesis
that biological feedbacks increase the Earth systems resistance or resilience to
perturbation, or increase its habitable range of conditions, and Lentons framing of
terms in this way has been helpful. This hypothesis can be useful as an organizing
concept, as a means of focusing attention, and as a spur to inquiry, and years ago I
said the same thing about its precursor, the homeostatic Gaia hypothesis (Kirch-
ner, 1989). Hypotheses like these, as questions, motivate us to look at mechanismsin the Earth system and ask what their consequences are for its regulatory behavior.
But we should be careful not to confuse the question with the answer; it is the
journey (the process of inquiry) that is helpful here, not the putative destination (a
proclamation that the hypothesis is or isnt proven). What we should expressly not
8/2/2019 2003 62 Gaia Conjectures
4/25
24 JAMES W. KIRCHNER
do is tally up the examples for and against, and presuming that there were more
and better examples in favor consider the hypothesis to be proven and apply it
as a generalization, despite the exceptions. That would convert a perfectly fertile
hypothesis into a sterile orthodoxy.
Science is a process, a journey toward a better collective understanding. It is not
a mere contest, in which all one needs to do is to declare a winner and a loser. Thus
the most useful answer to a question such as, Do biological feedbacks increase
the Earth systems resistance to perturbation? is neither Yes, they do nor No,
they dont. The most useful answer instead is, These mechanisms do, and those
mechanisms dont, and here are their implications for how the system functions.
And here are all the other things we need to figure out as a consequence.
Particularly given the conceptual plasticity of the Gaia concept, one should be
cautious about even a well-meaning search for confirmatory examples. The risk is
not that such an effort might fail, but rather that it will always succeed through
open-ended re-interpretation of the theory, the observations, or both. It is worth
recalling the impact of other theories that invited similar re-interpretation of obser-vations, and that were similarly heralded as fundamental breakthroughs in human
knowledge some time ago:
These theories appeared to be able to explain practically everything that hap-
pened within the fields to which they referred. The study of any of them
seemed to have the effect of an intellectual conversion or revelation, opening
your eyes to a new truth hidden from those not yet initiated. Once your eyes
were thus opened you saw confirming instances everywhere: the world was
full of verifications of the theory. Whatever happened always confirmed it.
Thus its truth appeared manifest; and unbelievers were clearly people who
did not want to see the manifest truth . . . (Popper, 1963).
Watching from outside, I have the sense and I want to emphasize that thisis just my subjective impression that some of those involved with Gaia theory
are caught in the uneasy tension between two impulses. The first impulse is to
interpret the history of the Earth as an epic tale in which the organisms play heroic
starring roles. The second impulse, driven by a broad-minded curiosity, is to probe
the complexities of the Earth system puzzle wherever they may lead. I have the
sense (again subjectively) that over time, the second impulse is having a growing
influence on the development of the Gaia concept. In my view this is all for the
good.
3. Gaia and Planetary Regulation
Just as LW caricature Volks critique, they caricature mine as well, saying Kirch-
ner suggests that some Gaia theorists are still arguing for universal or overwhelm-
ing negative feedback as a feature of the Gaia system, as if this were a prerequisite
for regulation. I did not say that Gaias proponents argue for universal or
8/2/2019 2003 62 Gaia Conjectures
5/25
GAIA: CONJECTURES AND REFUTATIONS 25
overwhelming negative feedback. What I said was precisely this: Gaias propo-
nents appear to view the Gaia hypothesis as combining elements of what I have
termed Homeostatic Gaia (i.e., biologically mediated feedbacks stabilize the
global environment) and a qualified form of Optimizing Gaia (i.e., biological
modifications of the environment make it more suitable for life) (Kirchner, 2002).
That is, if anything, a highly conservative interpretation of the written record, as
the following examples illustrate:
The notion of the biosphere as an active adaptive control system able to main-
tain the Earth in homeostasis we are calling the Gaia Hypothesis. (Lovelock
and Margulis, 1974, titled Atmospheric homeostasis by and for the biosphere:
The Gaia hypothesis).
The Gaia hypothesis . . . postulates that the climate and chemical composition
of the Earths surface are kept in homeostasis at an optimum by and for the
biosphere. (Lovelock and Watson, 1982).
Planetary life must be able to regulate its climate and chemical state . . . thegreater part of our own environment on earth is always perfect and comfort-
able for life . . . Through Gaia theory, I see the Earth and the life it bears as a
system, a system that has the capacity to regulate the temperature and com-
position of the Earths surface and keep it comfortable for living organisms
(Lovelock, 1988).
The Gaia hypothesis, when we introduced it in the 1970s supposed that the
atmosphere, the oceans, the climate, and the crust of the Earth are regulated at
a state comfortable for life by and for the biota. Specifically, the Gaia hypoth-
esis said that the temperature, oxidation state, acidity, and certain aspects of
the rocks and waters are at any time kept constant, and that this homeostasis is
maintained by the organisms at the Earths surface. It is important to recognizethat the Gaia hypothesis so stated is wrong . . . Through Gaia theory I now
see the system of the material Earth and living organisms on it, evolving so
that self-regulation is an emergent property. In such a system active feedback
processes operate automatically and solar energy sustains comfortable condi-
tions for life. The conditions are only constant in the short-term and evolve
in synchrony with the changing needs of the biota as it evolves (Lovelock,
1995).
The Gaia theory proposes that organisms contribute to self-regulating feed-
back mechanisms that have kept the Earths surface environment stable and
habitable for life (Lenton, 1998).
The Gaia hypothesis of Lovelock states that life regulates Earths functioningfor its own benefit, maintaining habitable, or even optimum conditions for life
(Kleidon, 2002).
The Gaia theory proposes that the Earth system self-regulates in a habitable
state (Lenton, 2002).
8/2/2019 2003 62 Gaia Conjectures
6/25
26 JAMES W. KIRCHNER
From these three decades of re-statements of what Gaia means, one can discern
both enduring themes and important changes over time. The most important change
is that the notion of an Earth regulated by and for the biosphere has generally been
abandoned, nowhere more explicitly than in Lovelocks 1995 revisions to his 1988
book. The two enduring themes reflected in the passages quoted above, but more
so in the surrounding body of work on Gaia are that biological feedbacks con-
tribute to the regulation of the environment, and that they help to maintain habitable
or comfortable conditions for life (which presumably means the life forms that
are dominant under those conditions). I agree that such feedback processes exist,
and I have consistently said so in print (Kirchner, 1989, 1991, 2002). I have also
consistently argued that we need to pay more attention to cases in which biological
processes destabilize the environment or make it less comfortable for life.
4. Biotic Destabilization of Earths Environment
Think about it: in the Gaia literature, where can one find detailed elucidations of
the biologically mediated feedbacks that are thought to undermine the stability
of the Earths climate (e.g., Lashof, 1989; Lashof et al., 1997; Woodwell et al.,
1998; Cox et al., 2000)? When have Gaias advocates proposed a biologically
mediated positive feedback mechanism? The one example that I can think of is
the DMS hypothesis, but that was originally proposed when DMS appeared to act
as a global thermostat (Charlson et al., 1987); only later did it become apparent that
this thermostat is hooked up backwards, serving to make the Earth cooler when
it was cool and warmer when it was warm (Legrand et al., 1988, 1991; Kirchner,
1990; Watson and Liss, 1998). In Gaian interpretations of the history of the Earth,
periods of rapid change are often attributed to external perturbations rather thaninherent instabilities in the Earth system. Likewise, destabilizing feedback is often
presented as an aberration that arises during the breakdown of regulatory mecha-
nisms (e.g., Lovelock and Kump, 1994), rather than an intrinsic characteristic of
many biologically mediated processes. Thus it seems to me that the Gaia literature
selectively emphasizes biologically mediated processes that produce stabilizing
feedbacks rather than destabilizing ones.
In response to my summary of biological feedbacks affecting global warming,
drawn from reviews by Lashof (1989), Lashof et al. (1997) and Woodwell et al.
(1998), LW note that, When considering future global warming, it is most impor-
tant to know the overall sign of feedback, which cannot simply be deduced from
how many feedbacks are positive and how many are negative (italics in original). I
agree. But if LW are familiar with this literature, they should be aware that Lashof
(1989) has answered precisely this question, by estimating the strength of a large
number of biologically mediated climate feedbacks (including the negative feed-
back resulting from direct fertilization of vegetation by CO2). Lashof concluded
that the overall sign is positive, with the net effect of amplifying the sensitivity of
8/2/2019 2003 62 Gaia Conjectures
7/25
GAIA: CONJECTURES AND REFUTATIONS 27
the Earth system to anthropogenic climate change. In Lashofs analysis, the overall
feedback gain for biologically mediated climate feedbacks is 0.080.44, compared
to an estimated gain of 0.170.77 for the sum of the water vapor, cloud albedo, and
snow albedo feedbacks. The uncertainties in any such calculation are substantial,
and Lashofs analysis may not be the final word on the subject, but it is currently the
best available quantitative summary of the best available data. As such, Lashofs
analysis is an empirical refutation of Lentons hypothesis that biological feedbacks
make the Earth system more resistant to perturbation.
Although the available evidence indicates that biological feedbacks will am-
plify global warming, the buildup of greenhouse gases in the atmosphere has been
slowed by biological uptake of anthropogenic CO2, as LW point out. But again it
is important to estimate the quantitative strength of this feedback. Anthropogenic
carbon emissions have raised the concentration of CO2 in the atmosphere from a
pre-industrial concentration of roughly 275 ppm to its present value of roughly
370 ppm, an increase of approximately 35%. The best available estimates indicate
that this increase in atmospheric CO2 has prompted an increase in the net rateof carbon uptake into the terrestrial biosphere of roughly 2 or 3 Pg C per year
(Schimel et al., 2001), whereas gross primary productivity and gross respiration of
the terrestrial biosphere are each roughly 120 Pg C per year (Schlesinger, 1997).
Thus the terrestrial biospheres gross primary productivity has increased (or its
gross respiration rate has decreased) by only 1.72.5%, in response to a 35%
increase in atmospheric CO2. Viewed in these quantitative terms, the coupling
between atmospheric CO2 and carbon uptake by the biosphere is weak, consistent
with Lashofs (1989) estimate of a negative feedback gain of only 0.02. From an
ecological perspective this is not surprising, since CO2 is not a limiting nutrient in
many ecosystems.
Nor does this picture change when uptake by the ocean is included. Gross car-bon uptake by the ocean is roughly 90 Pg C per year (Schlesinger, 1997), and has
increased by roughly 1.8 Pg C per year, or 2%, again in response to anthropogenic
increases in atmospheric CO2 totaling 35% (Schimel et al., 2001). Thus both ter-
restrial and oceanic uptake are only weakly coupled to atmospheric CO2. Note in
particular that whereas Gaia theory would predict that increases in atmospheric
CO2 should have a much stronger effect on terrestrial ecosystem uptake (which
is biologically mediated) than on ocean uptake (which is primarily abiotic), the
available data indicate that both are about equally insensitive to atmospheric CO2concentrations.
The Gaia hypothesis holds that the composition of the atmosphere is tightly reg-
ulated by biologically mediated feedbacks. Yet anthropogenic releases of carbon
have pushed atmospheric CO2 far beyond the range of concentrations seen on Earthfor the last 400,000 years, and this anomalously high concentration has persisted
for much longer than the mean lifetime of CO2 in the atmosphere. If atmospheric
CO2 were tightly regulated by the Earth system, the increase in atmospheric CO2could have been prevented by adjustments of only 34% in the gross fluxes between
8/2/2019 2003 62 Gaia Conjectures
8/25
28 JAMES W. KIRCHNER
the atmosphere, the oceans, and the terrestrial biosphere. The fact that this has not
happened demonstrates that atmospheric CO2 is not tightly regulated by the Earth
system, even though CO2 is an important controller of Earths climate, and even
though CO2
participates directly in the most fundamental processes of life. Thus
the failure of the Earth system to tightly regulate atmospheric CO2, at least on
human timescales, is another empirical refutation of the Gaia hypothesis.
The Vostok ice core record shows that to the extent that the Earth system
regulates CO2, CH4, and DMS in the atmosphere, all three of these planetary ther-
mostats are hooked up backwards, functioning to make the Earth cooler during
glacial periods and warmer during interglacials (Petit et al., 1999). LW claim that
this behavior is consistent with Gaia theory, noting that many complex systems
undergo transitions between states in which positive feedback predominates. I
agree that complex systems often undergo transitions driven by positive feedback.
But it makes no sense to call such behavior regulation, just as it would make no
sense to say that a drunkard was regulating the path of his truck as it swerved
down the street, alternately smashing into parked cars on the right and the left. Tothe extent that biologically mediated feedbacks control CO2, CH4, and DMS, they
apparently destabilize Earths climate on timescales of 100,000s of years. Thus the
ice core data provide yet another empirical refutation of the Gaia hypothesis.
5. Is Gaia More Consistent with the Distant Past?
Perhaps recognizing the extent to which the current behavior of the Earth system
tends to contradict the Gaia hypothesis, LW note that, . . . we must be clear about
the timescale of concern. Gaia theory is concerned with over 3.5 billion years of
Earth history. Focusing on the coming centuries and the past 1 million yearsis unlikely to give a representative picture, especially if we happen to live at an
unusual time of transition (italics in original). I agree that we might live in a time
of transition, but I see no reason to believe that times of transition are anomalous.
The geological record of the distant past does not necessarily paint a less dynamic
picture than the present, particularly when we recognize that the geological record
often disguises abrupt fluctuations by averaging them out (and it does so more
strongly, the farther back we look). New high-resolution isotopic measurements
for the more distant past are now becoming available, and they reveal abrupt
fluctuations in temperature and atmospheric composition scattered throughout the
geologic record at least as far back as the Cretaceous (Arens and Jahren, 2000;
Jahren et al., 2001; Zachos et al., 2001) and the Permian (Retallack, 2002).
Referring to the hypothesis that negative feedback is stronger in the presence
of life, LW argue that, Just because it is falsified at present on relatively short
timescales does not mean it is falsified on longer timescales or throughout Earth
history. We need to keep in mind that the recent past is the part of Earth history
that we know the best, and the present is the only part of Earth history in which we
8/2/2019 2003 62 Gaia Conjectures
9/25
GAIA: CONJECTURES AND REFUTATIONS 29
can directly observe Earth system processes at work. To the degree that the Gaia
hypothesis is contradicted by recent Earth history, where our observational data are
relatively good, one is not entitled to assume that it will necessarily fare any better
in the distant past, where our data are relatively poor. Indeed, in isotopic records
stretching back at least to the Permian, peaks in paleotemperature correspond to
peaks in paleo-CO2 (Retallack, 2002). Thus, if CO2 is biologically controlled as
part of a global thermostat, the data suggest that this thermostat has been hooked
up backwards for at least the past 300 million years.
We also need to keep in mind that as we look farther and farther back in time, our
interpretation of the geological and palaeontological record becomes more specu-
lative and ill-constrained. Particularly because Gaia theory and geophysiology
invite open-ended re-interpretation (Kirchner, 1989, 1990), it is hardly reassuring
that Gaia might be able to explain the behavior of the Earth system in the distant
past, where the theory and the data are both sufficiently ill-constrained that they
could be re-interpreted to make them consistent with each other.
Finally, whether or not Gaia theory adequately explains the Earth systemsbehavior a billion years ago, the fact that it is so plainly contradicted by many
prominent feedbacks in the present-day Earth system ought to put to rest any notion
that Gaia theory provides the essential theoretical basis for the putative profession
of planetary medicine (Lovelock, 1986; see also Lovelock, 1991 and chapter 7 of
Lovelock, 1995). Indeed, given the pervasiveness of destabilizing biological feed-
backs, it is appropriate to question whether the Earth system has been stabilized by
biological feedback processes, or in spite of them.
6. Biological Feedback at the Limits of Habitability
There are ultimate constraints to the extent to which biologically mediated feed-
backs can destabilize Earths climate. A biologically mediated feedback mecha-
nism cannot drive the environment to extremes that would cause the extinction
of the organisms responsible for that same feedback (Kirchner, 2002). Thus, as
Lenton (2002) has put it, feedbacks involving life automatically tend to stabilise
habitable conditions, because they involve biotic effects that can (by definition)
only operate under habitable conditions (italics in original). This would seem to
be virtually a proof-by-definition of the Gaia hypothesis, and to some extent it is.
However, it is important to remember that the single term life encompasses a
vast diversity of organisms, and the range of conditions that are habitable for
some organism or another is very broad indeed encompassing, for example,
temperatures spanning over 60 C and oxygen concentrations spanning at least two
orders of magnitude. Thus, while a biologically mediated positive feedback cannot
push environmental conditions beyond the range that is habitable for the organ-
isms responsible for it, it can push conditions outside the habitable range for other
organisms, and thus drive them to extinction. This is not just a hypothetical pos-
8/2/2019 2003 62 Gaia Conjectures
10/25
30 JAMES W. KIRCHNER
sibility. It has happened in Earth history, nowhere more dramatically than during
the Proterozoic, when a biologically driven rise in atmospheric oxygen drove the
previously dominant life forms (anaerobic bacteria) to extinction, everywhere that
they were exposed to Earths atmosphere (though they survive today in anaerobic
refuges, such as sediments and our own digestive tracts).
Nonetheless, I agree with LW that at the limits of habitability (for the specific
organisms involved), biologically mediated feedbacks should generally be stabi-
lizing. A good example of such a self-limiting biotic feedback is the tendency
for populations to grow until they have depleted their environments of essential
resources (nutrients, water, food, light, etc.) so severely that their further growth is
constrained. This phenomenon, which has recently (and aptly) been termed biotic
plunder by Toby Tyrrell of Southampton University, has been widely recognized
since at least the time of Malthus:
Among plants and animals the view of the subject is simple. They are all
impelled by a powerful instinct to the increase of their species, and thisinstinct is interrupted by no reasoning or doubts about providing for their
offspring. Wherever therefore there is liberty, the power of increase is exerted,
and the superabundant effects are repressed afterwards by want of room and
nourishment . . . (Malthus, 1798).
Likewise the growth of populations can be constrained by the wastes that they
produce, as populations foul their own nests and thus make their environments
less suitable for themselves (Kirchner, 1989). This is just another case of biotic
plunder, in which the resource that has been plundered is the environments capac-
ity to handle wastes. If one organisms wastes are another organisms food, then
a symbiotic interaction between them may serve to relax those particular resource
constraints. In that case, the population continues to grow until the biotic plunderof some other resource (ultimately either light or another source of free energy)
leads to such severe resource depletion that further growth is impossible.
This behavior is clearly stabilizing, as Lenton (2002) points out: . . . any bi-
otic effect that pushes environmental conditions towards the boundaries of what
is tolerable to the responsible organisms will be stopped by self-limiting negative
feedback. But this is hardly a beneficent mechanism that creates favorable con-
ditions for the biota. Instead it acts as the ultimate constraint to the organisms
capacity to make their environment wholly uninhabitable. This kind of feedback
often achieves stability at the outer limits of habitability; in the language of 200
years ago, misery is the check that represses the superior power of population
and keeps its effects equal to the means of subsistence (Malthus, 1798). A good
example is the biotic plunder of nutrients from the surface ocean as plankton take
up nutrients, die, and sink (Volk, 2002). This biologically mediated process has
created a biological desert over most of the worlds oceans, constituting over half
of Earths surface area. Biotic plunder by plankton is limited only by the nutrient
starvation of the plankton themselves.
8/2/2019 2003 62 Gaia Conjectures
11/25
GAIA: CONJECTURES AND REFUTATIONS 31
These negative feedbacks at the outer limits of habitability might be termed
a form of regulation just as, in my whimsical metaphor above, the parked cars
along the side of the street could be said to regulate the path of the drunkards
careening truck as it smashes into them, and thus is deflected back into the road-
way. Semantically speaking this may be regulation, but an environment that is
regulated in such a manner is hardly one that is always perfect and comfortable
for life (Lovelock, 1988).
7. Gaia and Natural Selection
LW misrepresent my argument when they say, Kirchner raises a theoretical prob-
lem that many evolutionary theorists have had with Gaia, that of cheats . I have
never raised the problem of cheats; neither the term nor the concept appears
anywhere in any of my papers on Gaia. Since my argument has been so thor-
oughly misunderstood, let me briefly try again; for a more complete explanation,
see pp. 401403 of Kirchner (2002).
The environment is, by definition, that which is shared among organisms. To
the extent that a trait improves or degrades the environment, it will affect its carri-
ers and its non-carriers equally. For a trait to spread via natural selection, it must
give its carriers an advantage over its non-carriers. Therefore, to the extent a trait
improves the environment rather than the individual, and thus benefits its carriers
and its non-carriers equally, natural selection will have no effect on it.
This fact may offend our sense of fairness, but it is an inevitable result of how
natural selection works. If a trait benefits individuals, its carriers will fare better
than its non-carriers and produce more offspring. As a result, that trait will become
more common in each successive generation. By contrast, if a trait benefits theshared environment, it will benefit its carriers and non-carriers alike and they
will both produce equal numbers of offspring. Thus that trait will not become
more common in the gene pool over time. Therefore, natural selection will gen-
erally favor traits that benefit individuals, whether they enhance the environment
or degrade it.
That was my central argument for why one should not generally expect natural
selection to favor environment-enhancing traits (or, for that matter, environment-
degrading traits). As I also pointed out, this argument may not hold if the
environment is not shared alike between carriers and non-carriers (such as in
metapopulations, e.g., Kirchner and Roy, 1999). In such cases, it is theoretically
possible for natural selection to favor environmental altruism, that is, to favor
traits that benefit the environment at a cost to the individual. For this to occur,
at least two conditions must be met. First, the ecosystem must be partitioned into
a collection of local organism/environment assemblages which compete with one
another, leading to a form of group selection. Second, the distribution of altruistic
individuals must be patchy, such that altruists are more likely to find themselves
8/2/2019 2003 62 Gaia Conjectures
12/25
32 JAMES W. KIRCHNER
among other altruists, and thus are more likely than non-altruists to share in the
benefits of altruistic behavior (Sober and Wilson, 1998). These conditions can
arise in spatially patchy populations with limited dispersal, where organisms en-
vironmental effects are localized in the immediate vicinity. If these conditions are
met, then whether environmental altruism is favored will depend on the relative
strength of selection between groups (favoring altruism) versus selection among
individuals within groups (favoring selfishness). Thus the hypothetical possibility
of Gaian altruism is not inconsistent with natural selection, but the hard work has
not yet been done; one needs to show that the necessary conditions arise in the real
world, and one needs to show that inter-group selection is sufficiently strong. One
cannot simply assume that traits will be favored by evolution because they are good
for the environment or the group; as Sober and Wilson (1998) have put it, Many
people axiomatically assume that societies, species, and ecosystems have evolved
to function harmoniously. It is not easy to grasp the fragility of this assumption,
and once grasped, the lesson learned is often troubling. I think that LW and Volk
share my view that this assumption is particularly fragile in the case of Gaia. Inparticular, to the extent that Gaia is principally concerned with the atmosphere and
oceans (which cannot be readily partitioned among groups) and concerned with the
global biota (comprising organisms with conflicting needs) it is difficult to see how
global-scale altruism can evolve, even through group selection.
In some cases, of course, traits that benefit the individual may also coinciden-
tally enhance the environment. Nitrogen fixation in nitrogen-limited ecosystems is
one such example, as LW suggest. Natural selection will favor these traits, but it
will generally favor them to the degree that they benefit the individual, not the envi-
ronment. As evidence of this, consider that legumes have a special structures (root
nodules) that help them retain as much of their hard-won nitrogen as possible, thus
minimizing the benefit that they provide to their environment. The benefit to theenvironment comes from the nitrogen that inadvertently leaks out as a by-product
of nitrogen fixation.
Traits that benefit the individual may also degrade the environment; for exam-
ple, many organisms are highly evolved to sequester resources for themselves, even
at the cost of impoverishing their environments. Natural selection will promote the
spread of such traits if they are beneficial for individuals, even though their end
result is biotic plunder and widespread resource depletion. In human affairs, the
counterpart to this phenomenon is termed the tragedy of the commons (Hardin,
1968).
8. Feedback on Selection
If traits alter the environment in ways that affect their carriers and non-carriers dif-
ferently, then their environmental consequences can affect natural selection. This
is feedback on selection in Lentons terminology. Using LWs example of nitro-
8/2/2019 2003 62 Gaia Conjectures
13/25
GAIA: CONJECTURES AND REFUTATIONS 33
gen fixation, as nitrogen fixers become more common in an ecosystem, nitrogen
becomes more widely available, thus diminishing their fitness advantage over non-
nitrogen-fixers. Over time, this negative feedback will modulate the abundance of
nitrogen fixers and non-fixers in response to changes in nitrogen availability, as LW
point out.
As another example, consider the traits that enable organisms to hoard resources
for themselves, thus depleting their environments. As resource-hoarding organisms
become more common (and thus resources become scarcer), organisms that effi-
ciently hoard resources will gain an ever-growing advantage over those that do not.
This positive feedback on selection serves to turbocharge the process of biotic
plunder, with the end result being a highly depleted environment and organisms
that are highly efficient at hoarding resources.
It is important to recognize that whether or not natural selection is mediated
through environmental changes (that is, feedback on selection), it cannot and will
not systematically favor traits that confer a general environmental benefit. I think
that LW understand this, but I belabor the point because many others do not. Traitsthat are beneficial to the individual may also confer a broader benefit as a by-
product (in the case of nitrogen leakage by nitrogen fixers), as LW and Volk (1998,
2002) have both pointed out. Alternatively, they may degrade the environment as a
by-product (as with resource-hoarding). Natural selection will favor them equally
in either case (at least in the absence of group selection). If a trait gives its carriers
an advantage over its non-carriers it will spread by natural selection, whether it
enhances or degrades the environment as a result.
9. Environmental Enhancement and Natural Selection
Discussions of Gaia often feature the notion that organisms generally alter their
physical and chemical environments in ways that benefit them. LW misrepresent
my argument on this point, attributing to me the view that the observation that the
environment is remarkably well suited to life does not necessarily mean that the
predominant process has been for life to alter the environment (emphasis added).
In fact I argued that this does not at all mean that life has altered the environment to
its benefit. What I said was that claims such as rainforest vegetation influences its
climate to its own benefit . . . are semantically correct but mechanistically mislead-
ing, because they suggest that the environmental conditions have somehow been
adjusted to the needs of the organisms. Instead, it is more mechanistically accurate
to say that natural selection has made rainforest organisms dependent on rainforest
conditions, which are partly of their own making (emphasis added).
Rainforest vegetation flourishes in the damp of the rainforest, and transpiration
by the dense vegetation contributes substantially to that dampness. But rainforests
are notwet because rainforest vegetation likes it that way. Even when such notions
are semantically purged of their teleological overtones, they still utterly fail as
8/2/2019 2003 62 Gaia Conjectures
14/25
34 JAMES W. KIRCHNER
mechanistic explanations for why rainforests are wet. There is simply no mecha-
nism by which the affinity of rainforest vegetation for dampness could be translated
into a process of natural selection that makes the rainforest wetter. Instead, the
particular organisms that will thrive in a rainforest are those for whom its wet, dark
environment is advantageous, whether or not they contribute to those wet, dark
conditions.
It is crucial to keep in mind that natural selection acts primarily on individuals
(Goodnight and Stevens, 1997), and that there is an inherent asymmetry between
individuals consequences for their environment, on the one hand, and the environ-
ments consequences for individuals, on the other hand. How much is the rainforest
environment altered by the addition or removal of a single tree (not an entire tree
species, just a single individual tree)? A little bit, but not much. By contrast, the
survival of each and every tree depends entirely on the physical, chemical, and
biological characteristics of its local environment.
In other words: each individual of each species has only an incremental effect
on its local environment, but the characteristics of the local environment com-pletely determine which individuals survive there and which dont. Therefore the
process by which organisms become well-matched to their environments is over-
whelmingly one of natural selection of organisms that happen to conform to local
environmental conditions, not one of the environment being somehow adjusted to
the benefit of the organisms that happen to live there. Claims that life alters the
environment to its benefit, which appear widely in the Gaia literature including
in much of Lovelocks work, in Kleidon (2002) and at least three times in Lenton
(1998) are fundamentally misleading. Of course the species that we find in a
rainforest benefit from the rainforest environment; if they didnt, they wouldnt
survive there, and we would find other species there instead!
The key point that is missing from Gaian discussions of organisms and their en-vironments is that the composition of the biota is not fixed in advance. Ecosystems
are open systems; their boundaries are porous to continuous dispersal of organisms
from other ecosystems. Organisms are continuously subject to competition, not
only from other species in their ecosystem but from species dispersing from other
ecosystems as well. This process of competition determines the species composi-
tion of the ecosystem, and guarantees that the dominant organisms will be those
best suited to the local environment.
As anyone who has walked through a forest knows, the local environmental
characteristics the availability of water, the amount of light streaming through
the canopy, the slope and aspect of the terrain, the abundance of pathogens and
parasites, the nutrient levels in the soil, and so forth change continuously from
place to place. Each point on the terrain is bombarded by seeds and spores; some ofthese originate in the local environment and others are would-be colonizers from
other environments. Only a tiny fraction of these seeds and spores will survive
to reproduce themselves. On average, the ones that survive and thrive will be the
ones that are best suited to the environmental characteristics of the particular point
8/2/2019 2003 62 Gaia Conjectures
15/25
GAIA: CONJECTURES AND REFUTATIONS 35
on the terrain where they happen to land. It is primarily through this process of
natural selection among competing species (and secondarily through the much
slower process of evolutionary adaptation of the traits of individual species) that
the organisms found at any point on the landscape become well-matched with their
local environmental characteristics.
The local environment is also incrementally altered by each individual that
lives there. Thus, as the species composition at each location changes, so do the
net effects on the environment. Through time, this process of natural selection
leads to the self-assembly of a collection of organisms (metaphorically termed a
community) that can survive and thrive in the conditions that they and their co-
occurring species have created. This is hardly a Gaian notion; ecologists have been
studying this process for decades.
Thus I agree with LW that there is a feedback loop between an ecosystems
species composition and its environmental conditions; I have consistently said
so, and I know no ecologist or Earth scientist who thinks otherwise. I have also
consistently said that the cumulative effect of a collection of organisms on theirenvironment can be substantial. But membership in that collection of organisms
and thus the match between those organisms and their environment is over-
whelmingly determined by the causal arrow that points from the environment to
the biota, not the other way around. Each individual alters its environment only in-
crementally, and it does not necessarily do so in ways that enhance the environment
for its needs.
The same principles apply at the global scale. Although we refer to the
global environment in the singular, we are actually referring to a diverse col-
lection of ecosystems, encompassing a vast range of environmental conditions.
The processes of dispersal and colonization described above are constantly at play
across these environmental gradients. As environmental conditions change, theyshift the boundaries between different communities, and thus alter the predomi-
nance of membership in the global biota. Thus during ice ages we may have more
tundra, and during interglacials we may have more tropical forests. These shifts in
ecosystem boundaries in turn affect the global climate, for example by altering the
planets albedo, latent heat budget, and atmospheric composition. Thus the climate
and the biota co-evolve. As the global environment evolves, the composition of
the global biota will shift, because those organisms whose needs are well met by
the environment will flourish: more penguins during glacial times, perhaps, and
more toucans during interglacials. Therefore, Gaian notions that environmental
conditions on Earth evolve in synchrony with the changing needs of the biota as it
evolves (Lovelock, 1995) are, from a mechanistic standpoint, exactly backwards.
The environment is always well-suited to the needs of the predominant organisms,precisely because that is why those particular organisms are predominant.
In the case of the rainforest environment, LW propose that one can test whether
organisms are flourishing primarily because of their impact on the environment or
8/2/2019 2003 62 Gaia Conjectures
16/25
36 JAMES W. KIRCHNER
primarily because they have adapted to the environmental conditions that they have
partly created, as follows:
If alteration of the environment and the resulting feedback dominates over
adaptation we would expect the rainforest-climate system to be prone torapid transitions when sufficiently perturbed, e.g., a switch to an arid pas-
ture/desert state. If adaptation has been the main shaping factor, we expect no
such collapse. Some models have predicted catastrophic collapse . . . We wait
pessimistically to see how the real system behaves.
I agree with LW that the rainforest-climate system may collapse if it is extensively
disrupted. I think that we also agree, at a mechanistic level, about why it is vulnera-
ble to collapse: the rainforest environment is partly a by-product of dense rainforest
vegetation, and a shift to a drier climate could prompt a thinning of rainforest veg-
etation, a reduction in local recycling of transpired water through the atmosphere,
and an invasion of other species better suited to the drier environment. Species that
thrive in a drier environment will typically transpire less water (because they cant
afford to waste it), thus making the environment drier still. So LW and I agree
that feedback occurs here, but we disagree about how to characterize it. To me, it
makes little sense to describe the rainforest environment as having been altered to
suit the needs of rainforest vegetation, just as it would make little sense to describe
dryland vegetation as altering its environment to its benefit when it refrains from
transpiring too much water (even though one could claim that this helps to prevent
a transition to a rainforest environment, which would be fatal for it). Instead, I
think the only mechanistically sensible view is this: the particular organisms that
we find in the rainforest are those for whom the rainforest environment is suitable,
and the rainforest environment is partly the result of the organisms that live there.
Thus rainforest species thrive in a rainforest environment that is partly of their
own making. But rainforests are not wet because rainforest species like it that way;instead, it is precisely whether an organism thrives in those damp, dark conditions
that determines whether it is found in a rainforest or in some other environment
instead.
Proponents of Gaia theory do not readily acknowledge that its two central
principles first, that organisms stabilize their environments, and second, that
organisms alter their environments in ways that benefit them are mutually incon-
sistent with one another (Kirchner, 1989). Consider the rainforest as an example.
Because rainforests are wet, vegetation can grow densely in them. Thus rates of
transpiration (and subsequent re-precipitation) of moisture are high, and the canopy
is tall and dense; all of these factors make rainforests even more humid. This in turn
permits even denser vegetation, which makes the rainforest even more humid, and
so on. This positive feedback process cannot go on forever, because it is ultimately
constrained by the finite supply of sunlight. But it is a positive feedback process,
and it is therefore destabilizing, as illustrated by the possibility of a rapid spiraling
collapse, described in the paragraph above. The rainforest moisture feedback is
destabilizing (and thus violates the first principle of Gaia theory) precisely be-
8/2/2019 2003 62 Gaia Conjectures
17/25
GAIA: CONJECTURES AND REFUTATIONS 37
cause it involves environmental effects of dense vegetation that promote still denser
vegetation (consistent with the second principle of Gaia theory).
10. Evolution of the Gaia Hypothesis
Reiterating almost verbatim an argument that has been made by Gaias proponents
for at least a decade (see Lovelock, 1990), LW say reactions to Gaia in the scientific
community have been progressing through the three classic stages first it is obvi-
ously wrong, then there may be something in it, and finally we have known it all
along. It seems to me that the communitys reaction to Gaia has changed largely
because Gaia itself has changed. The early versions of Gaia made extravagant
claims, such as global homeostasis by and for the biosphere, that indeed were
obviously wrong, as Lovelock (1995) now agrees. By contrast, recent statements
to the effect that The Gaia theory proposes that the Earth system self-regulates
in a habitable state (Lenton, 2002) are so self-evident that indeed we all really
have known it all along weve known that Earth self-regulates (since there is no
external agent to do the regulating), and weve also known from the fossil record
that Earth has been in a habitable state for much of its history.
As I said in my first Gaia commentary, it invites confusion when the single term
Gaia is used to refer to observations that are self-evident and propositions that are
highly speculative. That confusion persists today. Just compare the first sentences
of Kleidon (2002) and Lenton (2002): does Gaia mean only that the Earth system
self-regulates in a habitable state, or does Gaia mean that life regulates Earths
functioning for its own benefit, maintaining habitable, or even optimum condi-
tions for life? Is Daisyworld merely a parable (Watson and Lovelock, 1983)
that was never intended . . . to be more than a caricature (Lovelock, 2000), oris Daisyworld a cybernetic proof of the Gaia hypothesis (Lovelock, 1983) that
implies that seventy-five years of neo-Darwinist science will need to be rewritten
(Lovelock, 2000)? If Gaias proponents cannot be clear and consistent about what
Gaia means, it is unrealistic to expect the broader community to intuit its meaning
on their behalf.
Like it or not, there is a legacy associated with the Gaia label, resulting from
the extravagant claims of the 1970s and 1980s and from the ensuing New Age
hype. Many scientists impressions of Gaia are inevitably colored by the outlandish
claims that were stridently publicized under the Gaia label in the early years (and
were not directly retracted until 1995). Gaias supporters decry this situation, and
imply that the communitys views on Gaia are outdated. But it is simply myopic
to expect otherwise. Beyond the narrow sphere of Gaia supporters, it is unrealistic
to expect the community to keep up with the continually shifting definitions of
what Gaia means. It is easy to understand why many in the Earth system science
community want nothing to do with the Gaia label, because of the connotations
that it carries.
8/2/2019 2003 62 Gaia Conjectures
18/25
38 JAMES W. KIRCHNER
Thus my own view is that if there is to be a future for Gaia theory, it will involve
Gaias supporters integrating themselves into the larger Earth system science com-
munity, and abandoning the notion that Gaia stands apart from (or perhaps even in
opposition to) that broader effort. I think this will also require abandoning the Gaia
label, for the reasons outlined above.
Since this is likely to be the last commentary that I will write on Gaia for some
time, I would like to take the liberty to suggest several ways forward for Gaia
theory, in the hope that it can play a productive role in the further development of
Earth system science.
First of all, it is important to move beyond Daisyworld. Daisyworld has served a
useful purpose in illustrating how simple feedback relationships can produce self-
regulating behavior without conscious control. For those without prior training
in systems analysis, this is an important lesson to learn, and Daisyworld teaches
it effectively. For those unaccustomed to systems thinking, a first glimpse of the
phenomenon of systems-level emergent behavior often comes as a revelation, and
permanently alters ones outlook. Daisyworld is now widely used as a teachingtool for precisely this purpose, and that is all for the good. My only reservation is
that I think that Daisyworld has almost nothing useful to say about how climate is
regulated on the real Earth. By this I do not mean simply that on the real Earth,
vegetation albedo is only weakly coupled to climate (Schneider, 2001), or that this
coupling, however weak, is probably a positive feedback, just the opposite of what
the Daisyworld model assumes (Kirchner, 1989, 2002). Rather, my concern is that
even as an abstract model, Daisyworlds generality and thus its applicability to
the real world has been exaggerated.
Daisyworld is widely interpreted as demonstrating that atmosphere-biosphere
systems will always self-stabilize near the biological optimum. Daisyworld indeed
does this, but only because of a specific assumption embedded in the model. In thereal world, natural selection will generally favor behaviors that benefit individual
organisms, whether they improve the environment or degrade it. But in Daisy-
world, the only behaviors that benefit individuals are those that also improve the
environment. Thus, Daisyworld to some extent assumes what it sets out to prove
(for a more detailed discussion, see Kirchner, 2002). Although many elaborate
Daisyworlds have been developed over the years, to my knowledge they all retain
this built-in bias from the original Daisyworld model.
This bias can be reversed by making a biologically plausible change in a single
parameter. This simple parameter adjustment converts Daisyworld into a bi-stable
system that strictly avoids the biological optimum, as I showed over a dozen years
ago (see Figure 3 of Kirchner, 1989). What is most interesting about this alterna-
tive Daisyworld is that when subject to perturbations, its climate would undergounstable, positive-feedback transitions between its two end-member states, which
is at least qualitatively consistent with the ice core data (Kirchner, 1989).
But instead Daisyworld has, perhaps inadvertently, been configured to be
pathologically stable. Daisyworld is a one-feedback model; there is only one en-
8/2/2019 2003 62 Gaia Conjectures
19/25
GAIA: CONJECTURES AND REFUTATIONS 39
vironmental variable and it is regulated by extremely strong feedback with the
simplest possible biosphere. Such a simple model necessarily exhibits simple
behavior. By contrast, on the real Earth many different environmental variables
are coupled simultaneously, through many different feedback relationships, with
a highly complex biosphere composed of organisms with diverse (and often in-
compatible) environmental requirements. Such a complex system can exhibit many
kinds of behavior that a simple Daisyworld model cannot. I fully understand the
reasons for making highly simplified models, but the first rule of model-building is
to preserve the core features of the system that one is trying to study. The problem
is that Daisyworld is intended to explore the stability properties of hypotheti-
cal atmosphere-biosphere systems, but the stability properties of a one-feedback
system like Daisyworld are fundamentally different from those of more complex
systems, let alone realistically complex systems.
When I say there is a need to move beyond Daisyworld, I am not advocating the
construction of equally heuristic models of ever-greater complexity. In order to be
anything more than illustrations of abstract concepts in systems theory, modelsmust be mechanistically plausible and quantitatively realistic in relation to the
system under study. Otherwise, with a handful of feedback loops and arbitrary coef-
ficients, almost any model can produce almost any behavior. It is also essential that
model behavior be comparable against real-world behavior at comparable scales
(Schneider, 2001); otherwise models cannot be tested against data. Building and
testing quantitatively appropriate models of real-world atmosphere-biosphere in-
teractions, although harder work than building hypothetical Daisyworlds, is likely
to yield more relevant insights for Earth system science. Recent modeling efforts
by Lenton and colleagues (Lenton, 2001; Lenton and von Bloh, 2001; Lenton and
Watson, 2000) are a step in the right direction.
Second, it is important to move beyond simply theorizing. Nothing would domore to promote Gaias acceptance in the broader community than a successful test
of an a priori prediction, for which the canonical example is Eddingtons confirma-
tion of general relativity in 1919. Lovelocks observation that Earths atmosphere
is maintained in extreme chemical disequilibrium was a palpable early success, but
by itself this simply demonstrated biologys pervasive influence on the chemistry
of the atmosphere (which in any case was already largely understood by biogeo-
chemists see Hutchinson, 1954), rather than the regulatory coupling envisioned
by Gaia theory. After several decades and several iterations of Gaia theory, there is
an urgent need for that theory to be tested against data. Unfortunately, Gaia theory
makes predictions that are abstract and qualitative, making them difficult to test in
the real world.
However, the simple models developed by Gaia theory might be testable againstlaboratory microcosms, which they somewhat resemble. At the AGU Chapman
conference on Gaia in 1988, I remember attendees asking why nobody had built
laboratory microcosm experiments to demonstrate that organisms could mean-
ingfully regulate their own environment, the way Gaia theory says they should.
8/2/2019 2003 62 Gaia Conjectures
20/25
40 JAMES W. KIRCHNER
Hamilton (1995) reiterated this challenge seven years later, and as of today (seven
years still further on) the challenge remains unanswered, to the best of my knowl-
edge. Unless one can demonstrate that Gaia theory can correctly predict how a
simple microcosm experiment will evolve under controlled laboratory conditions,
is there any reason to expect that Gaia theory can correctly predict how the global
atmosphere-biosphere system will evolve under real-world conditions?
When testing models of Gaia, it is important to test them against observations,
not against ones preconceptions of how the world ought to behave (Kirchner,
1989). For example, Lovelock has repeatedly touted the fact that strong envi-
ronmental coupling strongly stabilizes ecosystems in Daisyworld, even when the
model includes many interacting species. This, according to Lovelock (1995),
shows the superiority of Daisyworld models over conventional biological models:
What is it, then, that confers the great stability and freedom from cyclical
and chaotic behavior on the Daisyworld models? The answer is that in Daisy-
world the species can never grow uncontrollably; if they do, the environment
becomes unfavorable and growth is curtailed. Similarly, while daisies live,the physical environment cannot move to unfavorable states; the responsive
growth of the appropriate colored daisy prevents it. It is the close coupling of
the relationships which constrain both daisy growth and planetary temperature
that makes the model behave.
But the problem is that making the model behave, in Lovelocks view, entails
making the model behave placidly, whereas this is rarely how real-world ecosys-
tems behave. Indeed, the whole reason that ecologists began modeling ecosystems
in the first place was to explain the highly unstable behaviors that often char-
acterize population dynamics in the real world the boom and bust cycles, the
local extinctions, and so forth. These real-world population dynamics are an
empirical refutation, both of Daisyworld and of preconceptions that the naturalworld is necessarily a tranquil place, harmoniously regulated at delicately adjusted
equilibria.
Third, it is important to move beyond generalities about life and the global
environment. The Gaia literature is marked by numerous conjectures about how
life affects the global environment, but one needs to recognize that neither life nor
the global environment are monolithic. The seductively simple term life encom-
passes a vast array of organisms with diverse and often contradictory environmental
requirements. The equally seductive term the global environment encompasses a
complex spatial patchwork of environmental conditions spanning, at a single point
in time, more than 60 C in temperature, 7 units of pH, three orders of magnitude
in ambient pressure, and several orders of magnitude in the concentrations of many
biologically important chemical constituents and thats just the habitable envi-
ronment. Thus attempts to generalize about either life or the environment can be
fraught with conceptual difficulties.
Many of the proposed tests for Gaia are based on assessing whether Earths
surface environment would be different without life. Would an abiotic Earth be
8/2/2019 2003 62 Gaia Conjectures
21/25
GAIA: CONJECTURES AND REFUTATIONS 41
less resilient and resistant to perturbation (Lenton, 2002)? Would the environ-
mental conditions prevailing on an abiotic Earth be less conducive to biological
productivity (Kleidon, 2002)? These hypotheses have the advantage of being fairly
precise, but the disadvantage of being untestable in practice. They also have the
more serious disadvantage of being far removed from the most pressing issues
in Earth system science namely, how the climate system works and how it will
evolve in the future. We know that a planet without life would be radically different
from the one we all live on now. Thus I think Kleidon could be correct that gross
primary productivity is higher on the present-day Earth than it would be under the
environmental conditions that would prevail on an abiotic Earth (particularly with a
present-day biota). But what would that prove? One is tempted to say that it would
prove the Gaia hypothesis, but it would only prove one specific conjecture based
on Gaia and one with little relevance to understanding the current Earth system.
How would that conjecture, assuming it were proven, help us to understand the
phenomenon of biotic plunder? Likewise I think Lenton could be correct that the
present-day Earth is more resistant to perturbation than an abiotic Earth would be.But how would that help us to understand why, according to the best available
evidence, biologically mediated feedbacks are likely to amplify the effects of an-
thropogenic global warming? The urge to seek generalizations is always strong,
but here I think a more nuanced approach is needed.
Fourth, it is important to keep the dialogue civil. My criticism of Gaia over
the years has been pointed, but I have always tried to be precise, factual, and fair,
and I have never engaged in name-calling. I regret that Lovelocks response has
been to characterize me as a figure of the Inquisition (Lovelock, 1990) and to
characterize my critique as a polemic (Lovelock, 1995) and as sophistry, not sci-
ence (Lovelock, 2000), while refusing to acknowledge that my critique is available
in print, thus preventing his audience from drawing their own conclusions on theissues at hand. Scattered through the Gaia literature are gratuitous slams such as
Geologists have tried to persuade us that the Earth is just a ball of rock . . . and that
life is merely an accident, a quiet passenger that happens to have hitched a ride on
this rock ball in its journey through space and time. Biologists have been no better
(Lovelock, 1995). These kinds of comments are unfair and counterproductive, and
they should stop. To pretend that Earth scientists and biologists live in completely
separate worlds is to deny the existence of the entire field of biogeochemistry,
which predates Gaia by decades (Kamen, 1946; Hutchinson, 1950, 1954; see also
Gorham, 1991).
Lovelocks scorn for evolutionary biologists takes me aback: A paradoxical
saying goes, The measure of a scientists eminence is the length of time he, or she,
holds up progress in the field. If confirmation were needed of Charles Darwinseminence this measure provides it . . . zealous disciples of Darwin by taking his
words as if they were revelations, not just the thoughts of a scientist, are making
a creed for biology and so hindering its natural development . . . In such a climate
of almost religious intolerance it was hardly surprising that the Gaia hypothesis
8/2/2019 2003 62 Gaia Conjectures
22/25
42 JAMES W. KIRCHNER
never stood a chance (Lovelock, 1995). The biologists of my acquaintance are
not narrow-minded zealots. Most of them disagree with Gaia, but they have valid
reasons for doing so. Most biologists and Earth scientists are good people, and the
way they have been caricatured in the Gaia literature (particularly in books written
for the general public, who are in no position to know any better) is a disservice to
science.
Gaia has helped to raise interest in the whole-system view of Earth, and that
is all for the good. It has helped to introduce systems concepts like feedback and
emergence to a wider audience, and that is an important contribution as well. Yet
Gaia remains marginalized. One reason is certainly Gaias New Age connotations,
but I believe that a larger problem is that many are put off by the tone simulta-
neously defensive and self-congratulatory of much of the Gaia literature. Here
too, however, there appears to be recent progress. The new generation of Gaia
supporters appears to be less eager to antagonize the rest of the community, and I
hope that this heralds the beginning of a new trend.
Finally, it is important to stay focused on the things that matter. It should notmatter whether the current version of Gaia theory survives intact. It should not mat-
ter whether the Gaia label endures. It certainly should not matter who is perceived
to have won the Gaia debate; as I said early in this essay, science is a journey
of exploration, not a paltry contest. What should matter is that we all keep moving
forward, toward a clearer understanding of how the Earth system works.
In exploring big questions, where our comprehension is puny compared to our
incomprehension, the most treacherous obstacle is a misguided sense that we al-
ready know the answers. My greatest worry about Gaia is that it could lead to a
false confidence that once one accepts that biological systems are intrinsically self-
regulating, understanding the Earth system is easy. The only check against this kind
of facile complacency to return to where this commentary began is the scientificprocess of conjectures and refutations. It is essential to remember that theories are
no more than conjectures, and to resist becoming entranced by them. And when
the data contradict those conjectures, it is essential to see them for the refutations
that they are. This is not easy our first instinct is often to rescue the theory by
denying or reinterpreting the data but in the long run, it is the only way forward.
We have made considerable progress toward unraveling the complexities of the
Earth system, but what we know is still dwarfed by what we dont know. What
controls the magnitude and duration of glacial/interglacial climate cycles, and why
have they changed over time? Why has the anthropogenic increase in atmospheric
CO2 so far resulted in a global temperature increase of only 1C or so, rather than
the much larger rise that one would expect from the correlation between temper-
ature and CO2 in the ice core data (Kirchner, 2002)? What controls the patternsof extinction and diversification in the fossil record, and how are they linked to
changes in climate? Anyone can come up with a long list of unanswered questions
like these, and the Earth system no doubt holds many other important puzzles that
we have not even imagined yet.
8/2/2019 2003 62 Gaia Conjectures
23/25
GAIA: CONJECTURES AND REFUTATIONS 43
And it is our home. Mankinds impact on the atmosphere and the biosphere is
growing daily, and growing at an ever-faster rate. We need to understand the con-
sequences of these trends for the global environment and the biota. We also need
to understand how to react constructively to this predicament. These are crucial
matters, and it is crucial to be clear-headed about them. We must see things as they
are, not as we wish they were. Let us go forward, then, with all the dedication, skill,
insight and clear-headedness that we can marshal for the task.
Acknowledgements
I thank Steve Schneider for encouraging this dialogue, and I thank Tim Lenton,
David Wilkinson, and Tyler Volk for their contributions to it; I also thank B. A.
Roy, J. Harte, and A. H. Jahren for their comments on the manscript. This work
was completed while I was supported by the Miller Institute for Basic Research.
References
Arens, N. C. and Jahren, A. H.: 2000, Carbon Isotope Excursion in Atmospheric CO2 at the
Cretaceous-Tertiary Boundary: Evidence from Terrestrial Sediments, Palaios 15, 314322.
Charlson, R. J., Lovelock, J. E., Andreae, M. O., and Warren, S. G.: 1987, Oceanic Phytoplankton,
Atmospheric Sulphur, Cloud Albedo and Climate, Nature 326, 655661.
Cox, P. M., Betts, R. A., Jones, C. D., Spall, S. A., and Totterdell, I. J.: 2000, Acceleration of Global
Warming Due to Carbon-cycle Feedbacks in a Coupled Climate Model, Nature 408, 184187.
Goodnight, C. J. and Stevens, L.: 1997, Experimental Studies of Group Selection: What Do They
Tell Us about Group Selection in Nature?, Amer. Nat. 150, S59S79.
Gorham, E.: 1991, Biogeochemistry its Origins and Development, Biogeochemistry 13, 199239.Hamilton, W. D.: 1995, Ecology in the Large: Gaia and Ghengis Khan, J. Appl. Ecol. 32, 451453.
Hardin, G.: 1968, The Tragedy of the Commons, Science 162, 12431248.
Hutchinson, G. E.: 1950, Survey of Contemporary Knowledge of Biogeochemistry.-3. The Biogeo-
chemistry of Vertebrate Excretion, Bull. Amer. Museum Nat. Hist. 96, 1553.
Hutchinson, G. E.: 1954, The Biogeochemistry of the Terrestrial Atmosphere, in Kuiper, G. P. (ed.),
The Earth as a Planet, University of Chicago Press, Chicago, pp. 371433.
Jahren, A. H., Arens, N. C., Sarmiento, G., Guerrero, J., and Amundson, R.: 2001, Terrestrial Record
of Methane Hydrate Dissociation in the Early Cretaceous, Geology 29, 159162.
Kamen, M. D.: 1946, Survey of Existing Knowledge of Biogeochemistry. 1. Isotopic Phenomena in
Biogeochemistry, Bull. Amer. Museum Nat. Hist. 87, 110138.
Kirchner, J. W.: 1989, The Gaia Hypothesis: Can it Be Tested?, Rev. Geophys. 27, 223235.
Kirchner, J. W.: 1990, Gaia Metaphor Unfalsifiable, Nature 345, 470.
Kirchner, J. W.: 1991, The Gaia Hypotheses: Are They Testable? Are They Useful?, in Schneider,
S. H. and Boston, P. J. (ed.), Scientists on Gaia, MIT Press, Cambridge, Massachusetts, pp. 3846.
Kirchner, J. W.: 2002, The Gaia Hypothesis: Fact, Theory, and Wishful Thinking, Clim. Change
52, 391408.
Kirchner, J. W. and Roy, B. A.: 1999, The Evolutionary Advantages of Dying Young: Epidemiolog-
ical Implications of Longevity in Metapopulations, Amer. Nat. 154, 140159.
8/2/2019 2003 62 Gaia Conjectures
24/25
44 JAMES W. KIRCHNER
Kleidon, A.: 2002, Testing the Effect of Life on Earths Functioning: How Gaian is the Earth
System?, Clim. Change 52, 383389.
Lashof, D. A.: 1989, The Dynamic Greenhouse: Feedback Processes that May Influence Future
Concentrations of Atmospheric Trace Gases in Climatic Change, Clim. Change 14, 213242.
Lashof, D. A., DeAngelo, B. J., Saleska, S. R., and Harte, J.: 1997, Terrestrial Ecosystem Feedbacksto Global Climate Change, Ann. Rev. Energy Environ. 22, 75118.
Legrand, M., Feniet-Saigne, C., Saltzman, E. S., Germain, C., Barkov, N. I., and Petrov, V. N.: 1991,
Ice-core Record of Oceanic Emissions of Dimethylsulphide during the Last Climate Cycle,
Nature 350, 144146.
Legrand, M. R., Delmas, R. J., and Charlson, R. J.: 1988, Climate Forcing Implications from Vostok
Ice-core Sulphate Data, Nature 334, 418420.
Lenton, T. M.: 1998, Gaia and Natural Selection, Nature 394, 439447.
Lenton, T. M.: 2001, The Role of Land Plants, Phosphorus Weathering and Fire in the Rise and
Regulation of Atmospheric Oxygen, Global Change Biol. 7, 613629.
Lenton, T. M.: 2002, Testing Gaia: The Effect of Life on Earths Habitability and Regulation, Clim.
Change 52, 409422.
Lenton, T. M. and von Bloh, W.: 2001, Biotic Feedback Extends the Life Span of the Biosphere,
Geophys. Res. Lett. 28, 17151718.
Lenton, T. M. and Watson, A. J.: 2000, Redfield Revisited 1. Regulation of Nitrate, Phosphate, and
Oxygen in the Ocean, Global Biogeochem. Cycles 14, 225248.
Lenton, T. M. and Wilkinson, D. M.: 2003, Developing the Gaia Theory: A Response to the
Criticisms of Kirchner and Volk, Clim. Change 58, this issue.
Lovelock, J. E.: 1983, Daisy World: A Cybernetic Proof of the Gaia Hypothesis, Coevolution
Quarterly 38, 6672.
Lovelock, J. E.: 1986, Geophysiology: A New Look at Earth Science, in Dickinson, R. E. (ed.), The
Geophysiology of Amazonia: Vegetation and Climate Interactions, Wiley, New York, pp. 1123.
Lovelock, J. E.: 1988, The Ages of Gaia, W. W. Norton, New York.
Lovelock, J. E.: 1990, Hands Up for the Gaia Hypothesis, Nature 344, 100102.
Lovelock, J. E.: 1991, Gaia The Practical Science of Planetary Medicine, Gaia Books, London.
Lovelock, J. E.: 1995, The Ages of Gaia, W. W. Norton, New York.
Lovelock, J. E.: 2000, Homage to Gaia: The Life of an Independent Scientist, Oxford University
Press, Oxford.Lovelock, J. E. and Kump, L. R.: 1994, Failure of Climate Regulation in a Geophysiological Model,
Nature 369, 732734.
Lovelock, J. E. and Margulis, L.: 1974, Atmospheric Homeostasis by and for the Biosphere: The
Gaia Hypothesis, Tellus 26, 29.
Lovelock, J. E. and Watson, A. J.: 1982, The Regulation of Carbon Dioxide and Climate: Gaia or
Geochemistry, Planet. Space Sci. 30, 795802.
Malthus, T. R.: 1798, An Essay on the Principle of Population as it Affects the Future Improvement
of Society with Remarks on the Speculations of Mr. Godwin, M. Condorcet and Other Writers ,
J. Johnson, London.
Petit, J. R., Jouzel, J., Raynaud, D., Barkov, N. I., Barnola, J.-M., Basile, I., Bender, M., Chap-
pellaz, J., Davisk, M., Delaygue, G., Delmotte, M., Kotlyakov, V. M., Legrand, M., Lipenkov,
V. Y., Lorius, C., Pepin, L., Ritz, C., Saltzmank, E., and Stievenard, M.: 1999, Climate and
Atmospheric History of the Past 420,000 Years from the Vostok Ice Core, Antarctica, Nature
399, 429436.Popper, K. R.: 1963, Conjectures and Refutations; The Growth of Scientific Knowledge, Routledge
and Kegan Paul, London.
Retallack, G. J.: 2002, Carbon Dioxide and Climate over the Past 300 Myr, Phil. Trans. Roy. Soc.
London Series A 360, 659673.
8/2/2019 2003 62 Gaia Conjectures
25/25